1. Field of the Invention
The invention relates to an image forming apparatus of the electrophotographic type, and particularly to an image forming apparatus of the electrophotographic type in which the image quality is controlled so as to be always held to a predetermined level.
2. Description of the Related Art
Generally, it is known that the sensitivity of a human being to a color difference is very high. When the color difference .DELTA.E between images to be compared with each other is about 5 in an L*a*b* color space, for example, the images can be distinguished from each other irrespective of the observer and conditions. It is reported that, when .DELTA.E is about 3, many observers hardly distinguish a color difference (see D. H. Alman, R. S. Berns, G. D. Snyder, and W. A. Larsen, Performance Testing of Color-Difference Metrics Using a Color Tolerance Dataset, COLOR research and application, vol. 14, Number 3, June 1989).
From these facts, it will be seen that, when the target level of the image reproducibility is set to be within the color difference recognition limit of a human being, an image forming apparatus is required to produce a color difference as high as .DELTA.E=3 or less.
As well known in the art, however, a prior art image forming apparatus of the electrophotographic type cannot fulfil such a high requirement. This is caused by the following reason. In an apparatus of the electrophotographic type, an electrostatic phenomenon is used, and hence the image output state of the apparatus itself is changed in accordance with the conditions of the environment where the apparatus is placed, such as the temperature and the humidity, or by temporal deterioration of a photosensitive member, a developer, or the like, with the result that the image reproducibility is varied.
To comply with this, an image forming apparatus of the electrophotographic type usually employs a feedback control to maintain the image density to an optimum level. In a control method which is most usually used, specifically, the reproduction state of the density and the environmental conditions of the interior of the apparatus are monitored by using a density patch, an error with respect to the target density is obtained, and the error is multiplied by a feedback gain, thereby calculating a correction amount of a preset value of a control actuator.
As the density patch, used is a density patch of an unfixed toner image after the developing process, or that of an image which is formed on a record medium such as a paper sheet and which has undergone the fixing process. A density patch of an unfixed toner image is used because such an image can be produced and erased more easily than a transferred image or a fixed image which is formed on a sheet, and has a high correlation of density with a fixed image. By contrast, a density patch of a fixed toner image is used because such an image itself is obtained by the user as a final image form and the quality of the image can be evaluated in consideration of factors including variation factors in the transferring and fixing processes.
As the environmental conditions of the interior of the apparatus which are to be detected, frequently, the temperature and the humidity which conspicuously affect the electrostatic phenomenon peculiar to the electrophotographic type are detected. As the control actuator, the voltage applied to the charger, the exposure value, the development bias, and the like which affect the developing properties are often used.
For example, Japanese Patent Publications (Kokai) Nos. SHO 63-177176, SHO 63-177177, and SHO 63-177178 disclose a technique in which the development density is desirably controlled by changing the development potential. Japanese Patent Publication (Kokai) No. HEI 1-169467 discloses another technique in which a density patch of a toner image is measured and exposure conditions and development bias conditions are controlled, thereby obtaining a desired image density.
However, the optimum development potential is always affected by various external factors which cannot be controlled, such as the temperature, the humidity, and the accumulated copy number. Therefore, there arises a difficulty that the development potential, the exposure value, the development bias, and the like must be set in consideration of these factors. Although a density patch of an unfixed toner image is highly correlated in density with a fixed image, it is impossible to detect an influence exerted by such factors in subsequent processes or the transferring and fixing processes.
In a method of monitoring the density of a fixed image, as typically shown in Japanese Patent Publications (Kokai) Nos. SHO62-296669 and SHO63-185279, an image reading unit incorporated into the main unit of an apparatus is often used. However, the work of transferring an image which has been once output to the image reading unit so as to be again subjected to the reading operation must be conducted by the user. As a result, this method has a disadvantage that such a very cumbersome work must be conducted in a usual image management routine.
Japanese Patent Publication (Kokai) No. HEI4-55868 discloses a technique in which the density of a fixed image is online monitored via an optical fiber, and Japanese Patent Publication (Kokai) No. HEI7-168412 discloses another technique in which such a density is online monitored by dedicated detecting means. In these methods, a test sheet must be output frequently or at each time when the image management is to be conducted. This produces a problem in that the user must bear the cost of the sheets. Since test sheets are frequently output, moreover, the image forming speed of the apparatus is substantially lowered, thereby producing a problem in that the original productivity of image formation of the apparatus is lowered.
Japanese Patent Publications (Kokai) Nos. HEI4-204461, HEI7-225505, and HEI3-87859 disclose a technique in which means for detecting conditions such as the environment is disposed and operations such as charging, exposure, and development are controlled. In the electrophotographic type in which electrostatic processes are mainly conducted, however, sufficiently high control accuracy cannot be always attained because relationships between optimum preset values of the development potential, the exposure value, and the development bias are nonlinear.
Because of these reasons, in an image forming apparatus of the electrophotographic type, effects such as those of various environmental conditions, for example, a high temperature and humidity state or low temperature and humidity state, and those of temporal deterioration of a photosensitive member, a developer, or the like must be previously known. As an apparatus of a higher control performance is to be developed, data must be collected in detail over a wider range of conditions. Therefore, enormous development manhours are required.
Furthermore, a feedback gain which is determined as a result of expending such enormous development manhours is not always optimum because of differences among apparatuses and a wide variety of use conditions. Particularly, the effect of temporal deterioration on the image density is largely varied depending on the degree of deterioration of parts used in each of apparatuses and the manner of operating each apparatus by the user. Therefore, it cannot be said that the control performance for image density is kept perfect for a long term after shipment.
Recently, methods which use a fuzzy system or a neural network system have been proposed as shown in Japanese Patent Publications (Kokai) Nos. HEI4-319971, HEI4-320278, and HEI5-207275.
In the proposed methods, however, only the control accuracy is improved by using a feature of a fuzzy system or a neural network system that it can cope with a case where the input and output relationship is complicated and nonlinear. Consequently, such methods are substantially useless for solving the above-discussed problems in that a large amount of data must be collected, that enormous development manhours are required, and that the control performance of each apparatus for the image density for a long term after shipment cannot be ensured.
In a fuzzy system, a membership function must be tuned by the engineer, and, in a neural network system, the learning operation may be automated but supervisory data for learning must be previously prepared by the engineer. In this way, actually, both the methods require considerably large development manhours.
Even in the case where a fuzzy system or a neural network system is used in which data of temporal deterioration are previously collected and which is configured in consideration of the collected data, there is a problem in that, when the input and output relationship is changed by actual temporal deterioration, differences among apparatuses, or replacement of parts, it is impossible to automatically cope with such a change. In other words, even when a fuzzy system or a neural network system is used, the control performance of each apparatus for the image density for a long term after shipment cannot be ensured.